Genome

Genome Based Taxonomic Classification

Journal: Genome

Manuscript ID gen-2018-0072.R4

Manuscript Type: Article

Date Submitted by the 15-Dec-2018 Author:

Complete List of Authors: Paul, Bobby; School of Life Sciences, Manipal Academy of Higher Education, Bioinformatics Dixit, Gunjan; School of Life Sciences, Manipal Academy of Higher Education Murali, Thokur Sreepathy; School of Life Sciences, Manipal Academy of Higher Education,Draft Biotechnology Satyamoorthy, K.; School of Life Sciences, Manipal Academy of Higher Education , Biotechnology

Keyword: Bacterial taxonomy, Bioinformatics, Nucleotide identity, Nomenclature

Is the invited manuscript for consideration in a Special Not applicable (regular submission) Issue? :

https://mc06.manuscriptcentral.com/genome-pubs Page 1 of 17 Genome

1 Genome Based Taxonomic Classification

2 Bobby Paul, Gunjan Dixit, Thokur Sreepathy Murali, Kapaettu Satyamoorthy* 3 School of Life Sciences, Manipal Academy of Higher Education, Manipal 576104, INDIA 4 5 6 7 8 *Corresponding Author: Dr. K. Satyamoorthy 9 Address: School of Life Sciences, 10 Manipal Academy of Higher Education, 11 Manipal, Karnataka 12 INDIA 576104 13 Email: [email protected] Draft

1 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 2 of 17

1 Abstract 2 Bacterial populations are routinely characterized based on the microscopic examination, colony 3 formation and biochemical tests. However, in recent past, bacterial identification, classification and 4 nomenclature have been strongly influenced by genome sequence information. Advances in 5 bioinformatics and growth in genome databases has placed genome-based metadata analysis in the 6 hands of researchers who will require taxonomic experience to resolve intricacies. To achieve this, 7 different tools are now available to quantitatively measure genome relatedness within members of 8 the same species and genome-wide average nucleotide identity (gANI) is one such reliable tool to 9 measure genome similarity. A genome assembly showed gANI score of <95% at intra-species level 10 is generally considered indicative of a separate species. In this study, we have analysed 300 whole- 11 genome sequences belonging to 26 different bacterial species available in the Genome database and 12 calculated their similarity at intra-species level based on gANI score. At the intra-species level, nine 13 bacterial species showed less than 90% gANI and more than 10% of unaligned regions. We suggest 14 appropriate use of available bioinformatics resources after genome assembly to arrive at proper 15 bacterial identification, classification andDraft nomenclature in order to avoid erroneous species 16 assignments and disparity due to diversity at the intra-species level. 17

18

19 Key words 20 Bacterial taxonomy; Bioinformatics; Nucleotide identity; Nomenclature

2 https://mc06.manuscriptcentral.com/genome-pubs Page 3 of 17 Genome

1 Introduction 2 Accurate bacterial classification is essential to identify the potential genes to understand genetic 3 mechanisms, disease diagnosis and provide better treatment especially during infectious disease 4 spread, outbreaks and progression (Fricke and Rasko 2014; Croucher and Didelot 2015; Kubicova 5 and Provaznik 2016). Bacterial classification is based on microbial techniques such as isolation, 6 culture and staining to determine morphological and biochemical features. Over the years, in 7 addition to morphological characters, phenotypic and genotypic methods have gained much 8 importance in bacterial identification and classification schema (Donelli et al. 2013). With the 9 rapid advancements in DNA sequencing techniques, genome-based methods are expected to play 10 a major role in delineating the bacterial species at genome level and provide valuable clues towards 11 their identification (Hugenholtz et al. 2016). Recently, metabarcoding and metagenomics have 12 become robust tools to study bacterial and archaeal diversity by comparative genomic analysis 13 (Srivathsan et al. 2015; Bowers et al. 2017). Increased reliability, decreased costs and time for 14 whole-genome sequencing and assemblyDraft has led to an unprecedented amount of information that 15 can provide significant insights towards drug-repositioning and development of robust molecular 16 diagnostic platforms (Fournier et al. 2014; Baltrus 2016; Hahnke et al. 2016; Hugenholtz et al. 17 2016; Jansen van Rensburg et al. 2016; Gosiewski et al. 2017). 18 It has been reported that there are millions of bacterial species of which only a fraction have been 19 identified and characterized (Eichorst et al. 2007; Kim et al. 2014). The global microbial species 20 richness is estimated to be anywhere between 107 to 109 (Schloss and Handelsman 2004). Till 21 December 2017, there were 17947 bacterial species with validly published names in Taxonomy 22 browser (https://www.ncbi.nlm.nih.gov/taxonomy). The development of whole genome 23 sequencing methods has resulted in the availability of large quantities of genomic data in public 24 repositories (Cook et al. 2016; Mukherjee et al. 2017). Over the past 20 years, more than 124,311 25 bacterial genome sequences have been deposited in the Genome database. However, the sequence 26 information needs critical evaluation before deciphering species descriptions, delineating related 27 species and assigning them into a reference genome database. Currently, bacterial species 28 delineation is based on phylogenetic analysis of marker genes such as 16S rRNA, 23S 29 rRNA, rpoB, gyrB, dnaK, etc. (Konstantinidis and Tiedje 2005, Liu et al. 2012). However, this 30 approach takes into consideration evolution of single genes for grouping the species or delineating 31 a species in a genus. Genome-level dissimilarity can happen due to the cross-species genome

3 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 4 of 17

1 contaminations (Merchant et al. 2014; Mukherjee et al. 2015; Kryukov and Imanishi 2016), lateral 2 gene transfer (Gillings 2017), and insertion sequences (Vincent et al. 2016). In addition, nucleotide 3 substitutions, gene insertions and deletions, gene duplications and rearrangements, which have 4 significant impact on evolution of prokaryotic genome (Coenye et al. 2005), are also considered 5 as critical determinants while defining a species. Horizontal gene transfer (HGT) plays an 6 important role in genome adaptation and evolution. Acquisition of new genes by HGT allows for 7 habitat-specific adaptation and ultimately results in a subsequent divergence of species or 8 speciation (Ochman et al. 2000). Additionally, genomes of several symbiotic bacterial species 9 show the existence of smaller genomes, codon reassignments and extreme biases in genome 10 sequences (McCutcheon and Moran 2012). Hence, it is imperative to look at whole genome 11 information rather than at selected genes for arriving at relatedness between species in a given 12 genus. 13 To further add to the complexity, advances in bioinformatics and growth in genome databases 14 requires researchers with taxonomic experienceDraft to perform genome-based metadata analysis in 15 order to critically evaluate the sequence information of genomes before assigning them into a 16 reference genome database. Though, a conclusive quantitative measure of genome relatedness is 17 indeed difficult, different methods have been routinely used to measure the relatedness within 18 members of the same species. The average nucleotide identity (ANI) is one such reliable measure 19 aimed at long-standing, systematic and scalable bacterial species classification (Kim et al. 2014). 20 Calculation of genome-wide ANI (gANI) score usually involves the fragmentation of genome 21 sequences, followed by calculating the nucleotide similarity between sequences by pairwise 22 alignment based only on the aligned regions. This is widely acknowledged as a robust measure of 23 genome relatedness based on the studies carried out on 13,151 prokaryotic genomes assigned to 24 3032 species (Varghese et al. 2015). Therefore the present study was aimed at identifying the 25 genome level similarity between the strains of selected bacterial species based on the data garnered 26 from genome assemblies available at the public databases by using parameter such as gANI. 27 28 Materials and methods 29 To highlight the genome-wide diversity at the species level, we analyzed a total of 300 genome 30 assemblies retrieved from the Genome database (https://www.ncbi.nlm.nih.gov/genome/), which 31 comprised of 26 unique bacterial species (Table 1). The selection of bacterial species for the study

4 https://mc06.manuscriptcentral.com/genome-pubs Page 5 of 17 Genome

1 was based on two criteria: a) the organism must have whole genome sequence in the completed 2 stage and b) the organism should have only one chromosome with about 10-15 available genome 3 assemblies. The genome sequence(s) were retrieved in Fasta file format and features including 4 genome size, %GC, number of genes and proteins were derived in tabular format. Using Microsoft 5 Excel, we calculated the differences in genome size, G+C content, number of coding genes and 6 non-coding genes at species level for all the short-listed strains. To determine the genome-wide 7 average nucleotide identity (gANI) at species level, multiple genome assemblies of the same 8 species were aligned with each other. Jspecies (http://jspecies.ribohost.com/jspeciesws/), a 9 BLAST-based tool was used for the calculation of gANI score and genome coverage (Richter et 10 al. 2015). Using R statistical package, boxplots for each species were generated from pairwise 11 genome identity and coverage. The 16S rRNA and 23S rRNA sequences from the 10 12 Flavobacterium psychrophilum strains used in this study were extracted, aligned using ClustalW 13 and a Maximum Likelihood tree (500 bootstrap; Tamura-Nei distance model) was constructed 14 using MEGA X software (Kumar et al. 2018).Draft 15 16 Results 17 We analyzed a total of 300 genome assemblies representing 26 unique bacterial species retrieved 18 from the Genome database (Table 1). A total of 10-15 strains were selected for each bacterial 19 species for the analysis. With respect to genome features, no significant difference was found in 20 the G+C content among the different strains of 26 bacterial genome assemblies analyzed. 21 However, genome assemblies for a single species, Prochlorococcus marinus, showed 20% 22 difference in total GC% among the different strains. We also observed a difference of more than 23 one Mb in genome size and a difference of more than 1000 coding genes within the strains 24 belonging to three species namely Flavobacterium psychrophilum, Prochlorococcus marinus and 25 Pseudomonas fluorescens (Table 1). The number of coding and non-coding genes observed was 26 found to vary proportionally according to their genome size. 27 Comparative genomics is a fundamental key to understand the functional elements across the 28 genomes and genetic similarity or diversity. We calculated the pairwise genome identity scores 29 and percentage of aligned genomic regions between the genome assemblies belonging to the same 30 species. A boxplot showing pairwise genome comparison identity (gANI) and percentage of 31 aligned genome region was constructed for all the 26 bacterial species (Fig. 1). We found nine out

5 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 6 of 17

1 of the 26 bacterial species analyzed to comprise of several strains with a gANI score of less than 2 90% at intra-species level (between strains of the same species). The minimum gANI scores (Table 3 1) were found in strains belonging to Prochlorococcus marinus (64.92), Flavobacterium 4 psychrophilum (69.26), Buchnera aphidicola (69.64), Clostridium botulinum (70.09), Pasteurella 5 multocida (78.4), Pseudomonas fluorescens (79.36), Aeromonas hydrophila (85.68), 6 Stenotrophomonas maltophilia (86.85) and Fusobacterium nucleatum (89.83). 7 As an example, to explore the conventional phylogenetic analysis using single marker genes for 8 grouping the species or delineating a species in a genus, we performed further analysis of the 10 9 strains of Flavobacterium psychrophilum. The 16S rRNA and 23S rRNA regions were extracted 10 from the whole genome sequences and compared. The gANI score between two F. psychrophilum 11 strains namely Z1 and Z2 was shown to be 100% while, the gANI scores of these two strains were 12 highly dissimilar from other strains studied (Table 2). 13 A BLAST based sequence analysis of 16S rRNA region also showed that F. psychrophilum Z1 14 and Z2 strains showed only 93% identityDraft with the other eight strains of F. psychrophilum studied 15 while, the 16S rRNA sequence shared more than 95% identity with other Flavobacterium species 16 (F. sp. NW20 [99%], F. hauense [98%], F. subsaxonicum [97%]) and 99% identity with 16S rRNA 17 sequences from uncultured bacterial species. Sequence homology of 23S rRNA region from ZI 18 and Z2 showed only 94% identity at intra-species level while, these regions showed a maximum 19 identity of 96% with F. arcticum. There were no 23S rRNA sequences available for F. sp. NW20, 20 F. hauense and F. subsaxonicum and hence they were not compared. Further, we performed a 21 phylogenetic analysis of 16S and 23S rRNA sequences from the 10 strains of Flavobacterium 22 psychrophilum strains (Fig. 2). The 16S rRNA and 23S rRNA sequence analyses showed that Z1 23 and Z2 strains form a distinct cluster compared to other strains. 24 Our analysis identified 20 bacterial species which showed <90% alignment against strains of the 25 same species (Table 1). The lowest percentage of aligned regions was found in strains of 26 Prochlorococcus marinus (16.37), Clostridium botulinum (21.62), Buchnera aphidicola (24.42) 27 and Flavobacterium psychrophilum (25.41). The median gANI identity score was found to be less 28 than 95% in six bacterial species (B. aphidicola, C. botulinum, F. nucleatum, P. marinus, P. 29 fluorescens, S. maltophilia). This clearly suggested that at least 50% of the genome assemblies 30 belonging to above indicated six species have significant variations in their genomes. A similar 31 result was observed for analysis of trees which were plotted based on the whole genome sequences

6 https://mc06.manuscriptcentral.com/genome-pubs Page 7 of 17 Genome

1 for the six bacterial species (available in the NCBI Genome database; see 2 https://www.ncbi.nlm.nih.gov/genome/tree/170 for B. aphidicola) which showed distinct clusters. 3 Similarly, only 12 species (B. anthracis, C. ulcerans, G. bethesdensis, K. pneumoniae, M. avium, 4 M. bovis, M. gallisepticum, M. haemolytica, P. acnes, P. multocida, R. prowazekii, R. rickettsii) 5 showed the median genome coverage of above 90% indicating that genome assemblies of the 6 remaining 14 species (A. hydrophila, B. breve, B. longum, B. aphidicola, C. difficile, C. botulinum, 7 F. psychrophilum, F. nucleatum, L. delbrueckii, M. mycoides, P. marinus, P. fluorescens, R. 8 anatipestifer, S. maltophilia) analysed have large stretches of unaligned unique regions. These 9 unique genomic regions might code for strain specific genes that might play a significant role in 10 their adaptation and survival. 11 12 Discussion 13 Comparative genomics analysis plays a major role in understanding evolution at both species as 14 well as genus level. By comparing all theDraft genomes at intra-species level, we can now determine 15 the degree of difference in a particular species. The average nucleotide identity calculated from 16 pairwise alignment between two genomes has been proposed as a new measure for taxonomical 17 classification of bacterial species (Zhang et al. 2014; Yoon et al. 2017). A putative ANI cut-off 18 score of 95% or above is suggested to indicate related species while, a score below 95% usually 19 suggests a ‘new’ species (Figueras et al. 2014; Han et al. 2016). For instance, the pairwise genome 20 comparison between strains of Pectobacterium parmentieri resulted in >99% gANI, suggesting 21 they belong to the same species; in contrast, Pectobacterium parmentieri and Pectobacterium 22 wasabiae showed a gANI score of 94%, indicating that they belong to different species (Khayi et 23 al. 2016). 24 In this study, we showed nine out of the 26 bacterial species were with remarkable diversity in 25 genome content with a gANI score of less than 90% at intra-species level. The genome 26 diversity/indels might have happened during evolution to maintain the relationship between 27 genotypic-phenotypic features and to enable adaptation to ecological or host habitat. These may 28 impart large diversity at species level, which is challenging for bioinformaticians to characterize 29 based on the data generated from a wide range of microbes using culture-independent high- 30 throughput metagenomics experiments.

7 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 8 of 17

1 We also observed that in 20 different bacterial species, a pairwise alignment showed unaligned 2 regions contributed to more than 10% of the genome, which clearly indicated that these genomes 3 contained several uncharacterized genes. Reasons for such a vast amount of unaligned regions may 4 be due to plasmids found in bacterial species or contamination from host genome (Merchant et al. 5 2014). Studies indicate that the genetic diversity derived from the plasmids might have roles in 6 bacterial survival at extreme habitats (Dziewit and Bartosik 2014). Inclusion of extrachromosomal 7 genomic regions in pairwise alignment will again increase the extent of unaligned regions. Hence, 8 it appears to be better to avoid such extra chromosomal regions for gANI based taxonomical 9 classification. 10 Recently, several bacterial species have been reclassified based on the gANI analysis (Godoy et 11 al. 2003; Stropko et al. 2014; Lawson et al. 2016; Vincent et al. 2016; Kook et al. 2017; Loveridge 12 et al. 2017; Sant’Anna et al. 2017; Liu et al. 2018). In an earlier study, Biller et al. (2014) observed 13 remarkable sequence diversity among 27 different Prochlorococcus strains from diverse oceanic 14 regions with wide variation in their GC%Draft (30.8-50.6), and these strains were further grouped under 15 physiologically distinct phylogenetic clades. However, till date, only one (Prochlorococcus 16 marinus) well classified species is available under the genus Prochlorococcus in the Taxonomy 17 browser. Environmental factors are known to influence GC content and are suspected to be the 18 major reason for the differences in GC content observed among closely related species that differ 19 sharply in genome sizes (Wu et al. 2012). Host habitat also plays a major role in species delineation 20 (Basharat et al. 2018) and in this context, comparison of gANI scores in relation to geographical 21 localities from which the strains were isolated is considered critical. However, in our study, we 22 could not classify the strains based on geographical location since this information is not available 23 for the strains whose whole genome sequences were included in the study. 24 Advances in genome sequencing technologies and bioinformatics approach for genome assembly 25 and annotations have resulted in accumulation of several bacterial genomes in public databases. 26 However, this also has resulted in erroneous species assignments leading to further confusions. 27 Hence, the Genomic Standards Consortium (http://gensc.org) has developed a set of standards for 28 the bacterial genome analysis and reporting it into the public domain (Bowers et al. 2017). 29 However, utilising available simpler bioinformatics resources will also provide more insights into 30 genome assemblies. For instance, in addition to genome identity, genome coverage could also be 31 considered as a major factor for genome classification and nomenclature. Similarly, inclusion of

8 https://mc06.manuscriptcentral.com/genome-pubs Page 9 of 17 Genome

1 authentic and reference gene sequences from the ‘same’ species during sequence analysis will 2 provide better clues on taxonomic identity of the isolate. Further, Genomic Standards Consortium 3 recommendations for the generation, analysis and reporting of bacterial single amplified genomes 4 should be strictly adhered to before depositing microbial whole genome sequences in the 5 international nucleotide sequence databases. 6 7 Conclusion 8 Number of whole genome sequencing projects has surged at an unprecedented momentum and is 9 expected to gather pace since the current number of complete genomes represent only a tiny 10 fraction of the total number of bacterial species on this planet. Over the years, number of studies 11 have highlighted large scale genetic variations at intra-species level and this has led to 12 reclassification of several bacterial species (Stropko et al. 2014; Lawson et al. 2016; Vincent et al. 13 2016; Kook et al. 2017; Loveridge et al. 2017; Liu et al. 2018). This scenario might arise if the 14 species delineation is based only on Draft the traditional methodology followed for the species 15 identification or due to outsourcing the genome sequencing and assembly to other service 16 providers. A move towards implementation of bacterial classification schemes based upon whole 17 genomes has significant implications in taxonomic classification and maintain genome level 18 commonality across the species in genome databases. Therefore, genome based taxonomic 19 classification can be considered when there are sufficient whole genome sequences available in 20 the database. The present study showed that genome assemblies belonging to the nine bacterial 21 species (P. marinus, F. psychrophilum, B. aphidicola, C. botulinum, P. multocida, P. fluorescens, 22 A. hydrophila, S. maltophilia and F. nucleatum) have significant differences (<90% gANI) at 23 species level. To understand the reason for this genetic diversity, we further analyzed genome 24 assemblies of F. psychrophilum as an example. Phylogenetic analysis of 16S and 23S rRNA gene 25 sequences from 10 assemblies showed that F. psychrophilum Z1 and Z2 strains form a distinct 26 cluster compared to other strains and homology search showed that these two strains strains have 27 higher degree of sequence similarity at inter-species level than intra-species level. Hence, we 28 recommend that the following steps aimed at reducing the genome level diversity at species level. 29 1. Estimation of sequence homology of potential marker genes such as 16S rRNA, 23S rRNA, 30 RecA, RpoB, GyrB, IleS, FusA using BLASTN program to identify relatedness of the

9 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 10 of 17

1 organisms at inter and intra species level. Also, traditional phylogenetic analysis using 2 these orthologous genes will provide information on relationships between taxa. 3 2. Calculation of genome wide average nucleotide identity (gANI) and percent unaligned 4 genomic regions as a potential measure of intra-species similarity. 5 3. The GC content, which is similar across most of the bacterial genome assemblies within a 6 species, codon usage and average amino acid identity (AAI) analysis will further help in 7 more accurately classifying an organism. 8 4. We recommend that researchers should strictly follow the recent guidelines by the genomic 9 standard consortium for genome submission and use available computational resources to 10 reduce the huge genome variations within species level.

11 Acknowledgements 12 The authors would like to thank DBT-FIST, Government of India and Manipal Academy of 13 Higher Education, Manipal for the support and facilities provided. 14 References Draft 15 16 Baltrus, D.A. 2016. Divorcing strain classification from species names. Trends Microbiol. 24: 17 431–439. doi:10.1016/j.tim.2016.02.004. 18 Basharat, Z., Tanveer, F., Yasmin, A., Shinwari, Z.K., He, T., and Tong, Y. 2018. Genome of 19 Serratia nematodiphila MB307 offers unique insights into its diverse traits. Genome. 61:469- 20 476. doi:10.1139/gen-2017-0250. 21 Biller, S.J., Berube, P.M., Berta-Thompson, J.W., Kelly, L., Roggensack, S.E., and Awad, L., et 22 al. 2014. Genomes of diverse isolates of the marine cyanobacterium Prochlorococcus. Sci. 23 Data 1: 140034. doi:10.1038/sdata.2014.34. 24 Bowers, R.M., Kyrpides, N.C., Stepanauskas, R., Harmon-Smith, M., Doud, D., and Reddy, 25 T.B.K., et al. 2017. Minimum information about a single amplified genome (MISAG) and 26 a metagenome-assembled genome (MIMAG) of bacteria and archaea. Nat. Biotechnol. 35: 27 725-731. doi:10.1038/nbt.3893. 28 Coenye, T., Gevers, D., Van De Peer, Y., Vandamme, P., and Swings, J. 2005. Towards a 29 prokaryotic genomic taxonomy. FEMS Microbiol. Rev. 29: 147-167. 30 doi:10.1016/j.femsre.2004.11.004. 31 Cook, C.E., Bergman, M.T., Finn, R.D., Cochrane, G., Birney, E., and Apweiler, R. 2016. The 32 European Bioinformatics Institute in 2016: Data growth and integration. Nucleic Acids 33 Res. 44: D20–D26. doi:10.1093/nar/gkv1352. 34 Croucher, N.J., and Didelot, X. 2015. The application of genomics to tracing bacterial pathogen 35 transmission. Curr. Opin. Microbiol. 23: 62-67. doi:10.1016/j.mib.2014.11.004. 36 Donelli, G., Vuotto, C., and Mastromarino, P. 2013. Phenotyping and genotyping are both essential 37 to identify and classify a probiotic microorganism. Microb. Ecol. Health Dis. 24: 1–8. 38 doi:10.3402/mehd.v24i0.20105.

10 https://mc06.manuscriptcentral.com/genome-pubs Page 11 of 17 Genome

1 Dziewit, L., and Bartosik, D. 2014. Plasmids of psychrophilic and psychrotolerant bacteria and 2 their role in adaptation to cold environments. Front. Microbiol. 5: 596. 3 doi:10.3389/fmicb.2014.00596. 4 Eichorst, S.A., Breznak, J.A., and Schmidt, T.M. 2007. Isolation and characterization of soil 5 bacteria that define Terriglobus gen. nov., in the phylum Acidobacteria. Appl. Environ. 6 Microbiol. 73: 2708–2717. doi:10.1128/AEM.02140-06. 7 Figueras, M.J., Beaz-Hidalgo, R., Hossain, M.J., and Liles, M.R. 2014. Taxonomic affiliation of 8 new genomes should be verified using average nucleotide identity and multilocus 9 phylogenetic analysis. Genome Announc. 2: e00927-14-e00927-14. 10 doi:10.1128/genomeA.00927-14. 11 Fournier, P.E., Dubourg, G., and Raoult, D. 2014. Clinical detection and characterization of 12 bacterial pathogens in the genomics era. Genome Med. 6: 114. doi:10.1186/s13073-014- 13 0114-2. 14 Fricke, W.F, and Rasko, D.A. 2014. Bacterial genome sequencing in the clinic: Bioinformatic 15 challenges and solutions. Nat. Rev. Genet. 15: 49–55. doi:10.1038/nrg3624. 16 Gillings, M.R. 2017. Lateral gene transfer, bacterial genome evolution, and the Anthropocene. 17 Ann. N. Y. Acad. Sci. 1389: 20–36. doi:10.1111/nyas.13213. 18 Godoy, F., Vancanneyt, M., Martinez, M., Steinbüchel, A., Swings, J., and Rehm, B.H.A. 2003. 19 Sphingopyxis chilensis sp. nov., a chlorophenol-degrading bacterium that accumulates 20 polyhydroxyalkanoate, and transfer of Sphingomonas alaskensis to Sphingopyxis 21 alaskensis comb. nov. Int. J. Syst.Draft Evol. Microbiol. 53: 473–477. doi:10.1099/ijs.0.02375- 22 0. 23 Gosiewski, T., Ludwig-Galezowska, A.H., Huminska, K., Sroka-Oleksiak, A., Radkowski, P., and 24 Salamon, D., et al. 2017. Comprehensive detection and identification of bacterial DNA in 25 the blood of patients with sepsis and healthy volunteers using next-generation sequencing 26 method - the observation of DNAemia. Eur. J. Clin. Microbiol. Infect. Dis. 36: 329–336. 27 doi:10.1007/s10096-016-2805-7. 28 Hahnke, R.L., Meier-Kolthoff, J.P., Garcia-Lopez, M., Mukherjee, S., Huntemann, M., and 29 Ivanova, N.N., et al. 2016. Genome-based taxonomic classification of Bacteroidetes. Front. 30 Microbiol. 7: 2003. doi:10.3389/fmicb.2016.02003. 31 Han, N., Qiang, Y., and Zhang, W. 2016. ANItools web: a web tool for fast genome comparison 32 within multiple bacterial strains. Database (Oxford) 2016: 1–5. 33 doi:10.1093/database/baw084. 34 Hugenholtz, P., Skarshewski, A., and Parks, D.H. 2016. Genome-based microbial taxonomy 35 coming of age. Cold Spring Harb. Perspect. Biol. 8: a018085. 36 doi:10.1101/cshperspect.a018085. 37 Jansen van Rensburg, M.J., Swift, C., Cody, A.J., Jenkins, C., and Maiden, M.C.J. 2016. 38 Exploiting bacterial whole-genome sequencing data for evaluation of diagnostic assays: 39 Campylobacter species identification as a case study. J. Clin. Microbiol. 54: 2882–2890. 40 doi:10.1128/JCM.01522-16. 41 Kim, M., Oh, H.S., Park, S.C., and Chun, J. 2014. Towards a taxonomic coherence between 42 average nucleotide identity and 16S rRNA gene sequence similarity for species 43 demarcation of prokaryotes. Int. J. Syst. Evol. Microbiol. 64: 346–351. 44 doi:10.1099/ijs.0.059774-0. 45 Konstantinidis, K.T., and Tiedje, J.M. 2005. Towards a genome-based taxonomy for prokaryotes. 46 J. Bacteriol. 187:6258-64. doi:10.1128/JB.187.18.6258-6264.2005.

11 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 12 of 17

1 Kook, J.K., Park, S.N., Lim, Y.K., Cho, E., Jo, E., and Roh, H., et al. 2017. Genome-based 2 reclassification of Fusobacterium nucleatum subspecies at the species level. Curr. 3 Microbiol. 74: 1137–1147. doi:10.1007/s00284-017-1296-9. 4 Kryukov, K., and Imanishi, T. 2016. Human contamination in public genome assemblies. PLoS 5 One 11: e0162424. doi:10.1371/journal.pone.0162424. 6 Kubicova, V., and Provaznik, I. 2016. Use of whole genome DNA spectrograms in bacterial 7 classification. Comput. Biol. Med. 69: 298–307. doi:10.1016/j.compbiomed.2015.04.038. 8 Kumar, S., Stecher, G., Li, M., Knyaz, C., and Tamura, K. 2018. MEGA X: Molecular 9 Evolutionary Genetics Analysis across computing platforms. Mol. Biol. Evol. 35: 1547– 10 1549. doi:10.1093/molbev/msy096. 11 Lawson, P.A., Citron, D.M., Tyrrell, K.L., and Finegold, S.M. 2016. Reclassification of 12 Clostridium difficile as Clostridioides difficile (Hall and O’Toole 1935) Prévot 1938. 13 Anaerobe 40: 95–99. doi:10.1016/j.anaerobe.2016.06.008. 14 Liu, W., Li, L., Khan, M.A., and Zhu, F. 2012. Popular molecular markers in bacteria. Mol. Genet. 15 Microbiol. Virol. 27: 103–107. doi:10.3103/S0891416812030056. 16 Liu, Y., Lai, Q., and Shao, Z. 2018. Genome analysis-based reclassification of Bacillus 17 weihenstephanensis as a later heterotypic synonym of Bacillus mycoides. Int. J. Syst. Evol. 18 Microbiol. 68: 106–112. doi:10.1099/ijsem.0.002466. 19 Loveridge, E.J., Jones, C., Bull, M.J., Moody, S.C., Kahl, M.W., and Khan, Z., et al. 2017. 20 Reclassification of the specialized metabolite producer Pseudomonas mesoacidophila 21 ATCC 31433 as a member of theDraft Burkholderia cepacia complex. J. Bacteriol. 199: e00125- 22 17. doi:10.1128/JB.00125-17. 23 McCutcheon, J.P., and Moran, N.A. 2012. Extreme genome reduction in symbiotic bacteria. Nat. 24 Rev. Microbiol. 10: 13–26. doi:10.1038/nrmicro2670. 25 Merchant, S., Wood, D.E., and Salzberg, S.L. 2014. Unexpected cross-species contamination in 26 genome sequencing projects. PeerJ 2: e675. doi:10.7717/peerj.675. 27 Mukherjee, S., Huntemann, M., Ivanova, N., Kyrpides, N.C., and Pati, A. 2015. Large-scale 28 contamination of microbial isolate genomes by illumina Phix control. Stand. Genomic Sci. 29 10: 18. doi:10.1186/1944-3277-10-18. 30 Mukherjee, S., Stamatis, D., Bertsch, J., Ovchinnikova, G., Verezemska, O., and Isbandi, M., et 31 al. 2017. Genomes OnLine Database (GOLD) v.6: Data updates and feature enhancements. 32 Nucleic Acids Res. 45: D446–D456. doi:10.1093/nar/gkw992. 33 Ochman, H., Lawrence, J.G., and Grolsman, E.A. 2000. Lateral gene transfer and the nature of 34 bacterial innovation. Nature 405: 299-304. doi:10.1038/35012500. 35 Richter, M., Rossello-Mora, R., Oliver Glockner, F., and Peplies, J. 2015. JSpeciesWS: A web 36 server for prokaryotic species circumscription based on pairwise genome comparison. 37 Bioinformatics 32: 929–931. doi:10.1093/bioinformatics/btv681. 38 Sant’Anna, F.H., Ambrosini, A., de Souza, R., de Carvalho Fernandes, G., Bach, E., and 39 Balsanelli, E., et al. 2017. Reclassification of Paenibacillus riograndensis as a genomovar 40 of Paenibacillus sonchi: Genome-based metrics improve bacterial taxonomic 41 classification. Front. Microbiol. 8: 1–13. doi:10.3389/fmicb.2017.01849. 42 Schloss, P.D., and Handelsman, J. 2004. Status of the microbial census. Microbiol. Mol. Biol. Rev. 43 68: 686–691. doi:68/4/686 [pii]\r10.1128/MMBR.68.4.686-691.2004. 44 Srivathsan, A., Sha, J.C.M., Vogler, A.P., and Meier, R. 2015. Comparing the effectiveness of 45 metagenomics and metabarcoding for diet analysis of a leaf-feeding monkey (Pygathrix 46 nemaeus). Mol. Ecol. Resour. 15: 250–261. doi:10.1111/1755-0998.12302.

12 https://mc06.manuscriptcentral.com/genome-pubs Page 13 of 17 Genome

1 Stropko, S.J., Pipes, S.E., and Newman, J.D. 2014. Genome-based reclassification of Bacillus cibi 2 as a later heterotypic synonym of Bacillus indicus and emended description of Bacillus 3 indicus. Int. J. Syst. Evol. Microbiol. 64: 3804–3809. doi:10.1099/ijs.0.068205-0. 4 Varghese, N.J., Mukherjee, S., Ivanova, N., Konstantinidis, K.T., Mavrommatis, K., and Kyrpides, 5 N.C., et al. 2015. Microbial species delineation using whole genome sequences. Nucleic 6 Acids Res. 43: 6761–6771. doi:10.1093/nar/gkv657. 7 Vincent, A.T., Trudel, M. V., Freschi, L., Nagar, V., Gagne-Thivierge, C., and Levesque, R.C., et 8 al. 2016. Increasing genomic diversity and evidence of constrained lifestyle evolution due 9 to insertion sequences in Aeromonas salmonicida. BMC Genomics 17: 1–12. 10 doi:10.1186/s12864-016-2381-3. 11 Wu, H., Zhang, Z., Hu, S., and Yu, J. 2012. On the molecular mechanism of GC content variation 12 among eubacterial genomes. Biol. Direct 7: 2. doi:10.1186/1745-6150-7-2. 13 Yoon, S.H., Ha, S. min, Lim, J., Kwon, S., and Chun, J. 2017. A large-scale evaluation of 14 algorithms to calculate average nucleotide identity. Antonie van Leeuwenhoek, Int. J. Gen. 15 Mol. Microbiol. 110: 1281–1286. doi:10.1007/s10482-017-0844-4. 16 Zhang, W., Du, P., Zheng, H., Yu, W., Wan, L., and Chen, C. 2014. Whole-genome sequence 17 comparison as a method for improving bacterial species definition. J. Gen. Appl. 18 Microbiol. 60: 75–78. doi:10.2323/jgam.60.75. 19 Draft

13 https://mc06.manuscriptcentral.com/genome-pubs Genome Page 14 of 17

1 Table 1 Genome level differences for isolates of selected 26 bacterial species

Sl. Organism No. of Genome Predicted Noncoding G+C gANIb Aligned No. genome size (Mb) CDS genes content (%) Identity (%) regions (%) assemblies Min Max Min Max Min Max Min Max Min Max Min Max 1 Aeromonas hydrophila 10 4.52 5.27 4164 4902 165 269 60.5 62.0 85.68 100 66.25 99.90 2 Bacillus anthracis 15 5.22 5.22 4952 5513 128 371 35.4 35.4 99.95 100 99.18 99.97 3 Bifidobacterium breve 10 2.24 2.42 1820 1960 135 214 58.6 59.0 97.31 98.82 81.77 94.44 4 Bifidobacterium longum 13 2.26 2.83 1796 2443 127 266 59.4 60.3 93.46 98.95 56.53 90.45 5 Buchnera aphidicola 10 0.44 0.64 355 577 45 223 23.0 26.3 69.64 99.92 24.42 76.36 6 Clostridioides difficile 11 4.04 4.29 3505 4128 143 251 28.5 29.2 95.49 100 74.93 90.23 7 Clostridium botulinum 12 3.61 4.39 3072 3971 66 360 27.4 28.3 70.09 100 21.62 87.58 8 Corynebacterium ulcerans 10 2.43 2.60 2035 2291 96 360 53.4 54.4 90.08 99.99 88.65 99.78 9 Flavobacterium psychrophilum 10 2.68 4.14 2311 3442 18 192 32.4 39.0 69.26 100 25.41 99.61 10 Fusobacterium nucleatum 10 2.03 2.54 1826 2325 19 198 26.8 27.2 89.83 99.97 52.93 83.65 11 Granulibacter bethesdensis 10 2.69 2.74 2338 2423 77 113 58.5 59.2 91.39 100 92.17 99.83 12 Klebsiella pneumoniae 15 5.08 5.52 4767 5194 206 318 57.0 58.0 93.32 99.98 82.66 99.73 13 Lactobacillus delbrueckii 13 1.85 2.26 1547 1950 251 387 49.1 50.1 97.03 99.36 73.77 94.20 14 Mannheimia haemolytica 10 2.50 2.70 2263 2651 151 226 41.0 41.1 98.40 100 82.60 99.12 15 Mycobacterium avium 12 4.78 5.62 4150 4955 212 406 68.8 69.3 97.17 99.99 80.63 97.89 16 Mycobacterium bovis 15 4.33 Draft4.41 3895 4778 48 331 65.6 65.7 99.63 100 95.69 97.58 17 Mycoplasma gallisepticum 12 0.92 1.01 701 745 61 113 31.4 31.6 96.82 99.94 89.67 97.05 18 Mycoplasma mycoides 12 1.08 1.21 845 1051 35 164 23.9 24.0 94.32 100 56.88 76.08 19 Pasteurella multocida 14 2.07 2.41 1885 2202 89 255 39.7 40.4 78.40 99.99 78.49 99.72 20 Prochlorococcus marinus 11 1.64 2.68 1784 2997 48 139 30.8 50.7 64.92 96.58 16.37 93.38 21 Propionibacterium acnes 10 2.48 2.56 2257 2416 130 216 60.0 60.1 97.03 99.98 92.63 99.88 22 Pseudomonas fluorescens 14 5.84 7.11 5179 6256 167 302 58.8 61.3 79.36 99.96 52.06 98.73 23 Rickettsia prowazekii 10 1.10 1.11 830 871 46 95 29.0 29.0 99.85 99.99 95.28 96.77 24 Rickettsia rickettsia 11 1.25 1.27 1246 1272 263 281 32.4 32.5 99.53 100 95.70 97.78 25 Riemerella anatipestifer 10 2.15 2.42 1801 2238 68 270 35.0 35.2 95 99.99 74.79 98.50 26 Stenotrophomonas maltophilia 10 4.50 4.85 3964 4368 112 287 66.3 67.4 86.85 100 67.16 99.59 2

https://mc06.manuscriptcentral.com/genome-pubs Page 15 of 17 Genome

Table 2. Whole genome average nucleotide identity (gANI) between different strains of Flavobacterium psychrophilum. Values in square brackets show percentage of aligned genomic region between strains.

STRAIN Z1 Z2 950106-1/1 FPG3 FPG101 CSF259-93 V4-33 V4-24 V3-5 JIP02/86 Z1 * 100.00 [99.61] 69.29 [25.42] 69.27 [25.58] 69.29 [25.47] 69.30 [25.48] 69.29 [25.48] 69.30 [25.41] 69.29 [25.48] 69.29 [25.47] Z2 100.00 [99.62] * 69.27 [25.45] 69.28 [25.59] 69.27 [25.46] 69.27 [25.51] 69.27 [25.51] 69.27 [25.44] 69.27 [25.51] 69.26 [25.49] 950106-1/1 69.87 [34.79] 69.87 [34.79] * 99.25 [82.58] 99.78 [92.13] 99.84 [92.78] 99.97 [94.21] 99.96 [93.95] 99.97 [93.89] 99.93 [94.64] FPG3 70.42 [35.77] 70.43 [35.74] 99.19 [84.67] * 99.21 [87.51] 99.19 [87.41] 99.09 [85.41] 99.23 [84.65] 99.17 [85.32] 99.25 [87.34] FPG101 70.08 [34.25] 70.08 [34.25] 99.71 [90.24] 99.22 [83.45] * 99.89 [94.75] 99.65 [90.87] 99.74 [90.40] 99.71 [91.06] 99.83 [93.52] CSF259-93 70.07 [33.49] 70.07 [33.49] 99.76 [89.31] 99.06 [82.22] 99.78 [93.72] * 99.66 [89.98] 99.77 [89.39] 99.71 [90.01] 99.81 [93.03] V4-33 69.76 [34.85] 69.76 [34.84] 99.97 [94.94] 99.32 [83.63] 99.79 [93.19] 99.86 [93.90] * 99.96 [94.95] 99.97 [95.26] 99.94 [94.16] V4-24 69.74 [35.29] 69.75 [35.28] 99.97 [95.51] 99.31 [84.22] 99.81 [93.44] 99.88 [94.17] 99.98 [95.67] * 99.97 [95.55] 99.94 [94.46] V3-5 69.77 [35.36] 69.77 [35.36] 99.95 [95.16] 99.28 [84.22] 99.78 [93.47] 99.87 [94.11] 99.96 [95.58] 99.96 [95.00] * 99.93 [94.40] JIP02/86 70.36 [33.68] 70.36 [33.68] 99.82 [91.86] 99.18 [83.06] 99.81 [92.63]Draft99.87 [93.41] 99.75 [90.99] 99.83 [90.47] 99.82 [91.06] *

https://mc06.manuscriptcentral.com/genome-pubs Genome Page 16 of 17

Figure 1 Boxplot showing percentage genome identity (A) and coverage percentage (B) at intra-species level for the bacterial species studied. The boxes represent the 25th-75th percentile values with the median score while the circles indicate outliers.

347x175mmDraft (300 x 300 DPI)

https://mc06.manuscriptcentral.com/genome-pubs Page 17 of 17 Genome

Figure 2. Phylogenetic tree based on the nucleotide sequences of A)16S ribosomal RNA and B) 23S ribosomal RNA genes from 10 Flavobacterium psychrophilum strains selected in the study. The phylogentic tree was constructed by the Maximum Likelihood method, using the MEGA X Software.

513x186mm (300 x 300 DPI) Draft

https://mc06.manuscriptcentral.com/genome-pubs